Great Chart: US Term Premium vs. Business Cycles

Academics and economists have often decomposed the long-term bond yield of a specific country (i.e. US 10Y Treasury) into the sum of the expected path of real interest rate (r*) and the additional term premium, which compensate investors for holding interest rate risk. Two major risks that a bond investor typically face in the long-run are the change in supply of and demand for bonds and the uncertainty around inflation expectations. If the uncertainty increases, the market will demand a higher premium as a response. As the premium is not directly observable, it must be estimated using econometric models. For instance, a popular one that practitioners use is the one developed by Adrian, Crump and Moench (2013), who estimated fitted yields and the expected average short-term interest rates for different maturities (1 to 10 years, see data here).

As you can see, the term premium has been falling since 2009 and is currently negative at -51bps, which has not happened very often. Instead of having a positive term premium for long-term US debt holders carrying interest rate risk, there is actually a discount. The term premium for the 10Y reached an all-time low of -84bps in July 2016, at the same time that the yield on the Treasury reached a record low below 1.40%. However, there are also interest findings when we plot the ACM 10Y term premium with macroeconomic variables. If we overlay it with the US unemployment rate, we can notice a significant co-movement between the two times series. The jobless rate went down from 10% in Q3 2009 to 4.1% in March 2018, tracked by the term premium that fall from roughly 2.5% to -50bps in that same period. In other, it seems that the term premium follows the business cycles, trend lower in periods of positive growth and falling unemployment and rises in periods of contractions. Therefore, for those who are expecting a rise in the US LT yields in the medium term, driven by a reversion in the term premium, what does it mean for the unemployment rate going forward?

Chart. US Unemployment Rate vs. 10Y Term Premium

US Term Pr.png

Source: Reuters Eikon and Adrian et al. (2013)

 

Great Chart: Term Spread Differentials (US, Germany and Japan)

In this article, we define the term spread of a specific country by the difference between the long-term (10Y) and the short-term (2Y) sovereign yield, which is also referred as the yield curve. As we mentioned it in one of our previous Great Chart articles (here), empirical research has shown a significant relationship between the real economic activity of a country and the yield curve. In today’s edition, we chose to look at the historical developments of the term spread differentials, between the US and Germany and the US and Japan.

Over time, we notice that the term spread has some interesting co-movement with the exchange rate. For instance, between 2005 and 2017, a widening term spread differential between the US and Germany was favourable to the USD/EUR exchange rate (here), meaning that the Euro was appreciating when the US yield curve was steepening more significantly than the German one. However, we saw that the relationship between the two times series broke down in early 2017 and has actually reversed over the past 14 months (here). In other words, based on the current market levels, the 2Y10Y term premium in Germany offers 56bps more than the US. Hence, as the term structure in the US has flattened strongly relative to Germany (yield curve steepened from 50bps in July 2016 to 118bps), the US Dollar depreciated.

This chart shows the evolution of the term spread differentials – between US and Germany and between US and Japan – since 1985. We can observe a strong correlation between the two times series over the past 30 years, with the term spread differential against Germany trading at -57bps, its lowest level since June 2006, and at 42bps against Japan, its lowest level since June 2008, respectively. An interesting observation comes out when we look at the spread between the two TS differentials (US-Japan vs. US-DE), which simply comes back at looking at the cross term spread differential between Germany and Japan. At the exception of the year 1992, the DE-Japan TS differential has always traded between -1% and +1%, and is currently standing at the high of its long-term range. The TS differential currently trades at +1% on the back of a steepening German yield curve since the summer of 2016 (2Y10Y moved from 52bps in July 2016 to 119bps today). It it a good time to play the convergence between the two term structure, i.e going long the German 2Y10Y term spread and short Japan 2Y10Y? The risk of the trade is on Japan side, as shorting the 2Y10Y would imply a steepening yield curve with either the 2Y yield going down or the 10Y rising. With the current BoJ ‘yield curve control’ (YCC) policy, we know that a steepening yield curve in Japan is difficult for the time being, but it will be interesting to see where TS differentials stand in a couple of months.

Chart: Term spread Differentials – Japan and Germany vs. US (Source: Reuters Eikon)

Term Spreads ALl

Great Chart: Italy EPU Index vs. 10Y Bond Yield

The recent results in Italian’s election held on March 4th wasn’t really a surprise for market participants, with EURUSD barely moving (the pair is actually up 2.5 figures over the past week) and the 5Y CDS spread (vs. Germany) flat at around 92bps (here). According to the latest estimates, the populist Five-Star movement, created by comedian Beppe Grillo and led by its prime ministerial candidate Luigi di Maio, came in first individually capturing 32.7% of the votes. However, if we look at the coalitions results, the Center-Right coalition got 37% of the vote shares, with the alliance including the League with 17.4%, former prime minister Silvio Berlusconi’s Forza Italia (14%) and the Brothers of Italy (4.4%) and US with Italy (1.3%) parties. The disappointment was for the Democratic Party, which has governed Italy since 2013, as the Center-Left coalition captured ‘only’ 23% of the vote shares (much lower than the 27+% estimates, here), prompting former PM Matteo Renzi to step down as party leader. The FT published an interesting graphic lately, showing the geography of the electoral vote: Italy, the politically divided country (here). As you can see it, the Five-Star movement made the largest gain in the South (including Sardinia), in regions with the lowest per capita income.

Hence, following the election results, an interesting chart to watch in the weeks to come is the 10Y Bond yield vs. the Italy EPU index. As a reminder, the Economic Policy Uncertainty (EPU) index was developed by Baker, Bloom and Davis (2016) as a measure of economic policy uncertainty based on newspaper coverage frequency. The authors studied the evolution of political uncertainty since 1985 across countries (12 including the US) using leading newspapers that contain a combination of three of the target terms: economy, uncertainty and one or more policy-relevant terms (For the European EPU index, the author used two leading newspapers per country). Since its inception, the index has gained popularity in practice, measuring another form of market’s volatility or uncertainty. Baker et al. found that elevated political uncertainty has negative economic effects, which can potentially impact market prices.

This chart plots the EPU index versus the Italy 10-year bond yield. We can observe an interesting correlation between the two series. Since the financial crisis, it looks like LT sovereign yields have been rising when the EPU index increased ahead of a political or economic uncertain event. For instance, during the European debt crisis of 2010 – 2012, the EPU Index for Italy rose from 75 to over 200, while the 10Y yield skyrocketed from 4% to 7%. The financial meltdown in the Euro area was then halted after ECB Draghi’s “Whatever it takes to preserve the Euro” famous words at a global investment conference in London on 26 July, 2012.

As we mentioned in our previous posts, we don’t see any imminent risk for Italy, however a potential threat to investors would be a prolonged period of political instability. The question now is: can a rise in Italian LT yields in the next few months lead to a contagion to other peripheral countries’ bond yields (i.e. Spain or Portugal, here)?

Chart: Italy EPU Index (lhs) vs. 10 bond yield 

(Source: Eikon Reuters, policyuncertainty.com)

 

The Balassa-Samuelson Effect and The MEVA G10 FX Model

Abstract: In this study, we introduce Danske’s Medium Term FX Evaluation model (MEVA G10 FX), a framework that falls within the class of the Behavioural Equilibrium Exchange Rate (BEER) models. An important concept of the BEER model is that there is no prior theory for the choice of economic variables; hence, the choice of variables is based on economic intuition and data simplicity and availability.

Using two medium-term G10 FX drivers – a gauge of the Balassa-Samuelson effect and the terms of trade – we run a Fixed-Effect panel regression on the G10 currencies, using the US Dollar and the Euro as the base currencies.

PDF LINK ===========================> MEVA G10

EXCEL DATA LINK ====================> MENA FX – Quarterly Data

EXCEL (BALASSA – SAMUELSON) DATA ===> PennEffect

Results of our study (FX Q1 2018 spot rates were from mid-february) 

Great Chart: US Yield Curve vs. VIX (log, 30M lagged)

As a response to the recent surge in the market’s volatility (VIX), we saw lately an interesting chart that plots the 2Y10Y yield curve overlaid with the VIX (log, 30-month lagged). Even though we don’t necessarily agree with the fact that yield curves are a good predictor of recessions, we like to integrate it in our analysis as a supportive argument when presenting our outlooks as it summarizes a lot of information in a single chart. Previously, we presented the SP500 index versus the 2Y10Y yield curve (here), in which we emphasized that US equities can continue to rise (as the fundamental indicators) for weeks (2000) or months (2006/2007) despite a negative yield curve.

In this chart, we can notice another important factor, which is that the bull momentum in the equity market can persist even though market experiences an increase in price volatility (on an implied base). For instance, in the last two years of the 1990s (98/99), the VIX averaged 25%, 10 percent higher than in the last few years, while the SP500 was up 70% (the Nasdaq actually increased by 100% in the last quarter of 1999).

Hence, if we assume that the 25-year relationship between equity volatility and the business cycle holds on average, the constant flattening US yield curve over the past 2 years was suggesting a rise in the VIX.  The chart shows the persistent divergence between the two times series prior the sell-off; while the 2Y10Y had flattened by 200bps to 0.50% over the past couple of years, the VIX was averaging 10-12. The question now is: what to expect in the future for US equities, volatility and yields?

With the 10-year slowly approaching the 3-percent threshold, are US equities and volatility sensitive to higher long-term yields? As Chris Cole from Artemis pointed out in his memo Volatility and the Alchemy of Risk, there is an estimated 2tr+ USD Global Short Volatility trade (i.e. 1tr USD in risk parity and target vol strategies, 250bn USD in risk premia…). Can we experience another late 1990s period with rising LT yields, higher implied volatility without a global deleveraging impacting all asset prices?

In our view, it is difficult to see a scenario with rising LT yields combined with an elevated volatility (i.e. 20 – 25 %) without a negative impact on overall asset classes. Hence, if we see a persistent high volatility in the medium term as this chart suggests, the deleveraging in both bonds and equities by investment managers will kickstart a negative sell-reinforcing process, creating a significant sell-off in all asset classes with important outflows in the high-yield / EM investment world, hence leading to a repricing of risk.

Chart. US 2Y10Y Yield Curve vs. VIX (log, 30M lagged) (Source: Eikon Reuters)

USYield vs VIX

Great Chart: Relative Implied Volatility – VIX/RVX ratio

For each investor, there are several ways of measuring the market’s temperature. For instance, former Fed chairman Alan Greenspan would look at the 10-year US yield, some investment managers will simply look at the VIX and currency traders will tend to watch the moves on the Japanese Yen, especially against the US and Australian Dollar (see AUDJPY and SP500 correlation here). We know empirically that a sudden move on the Yen (JPY appreciates relative to other currencies) is usually accompanied with an equity correction and hence an increase in the implied volatility. Even though we hear a lot about the VIX measure, we also need to pay attention to the implied volatility surface, presenting skew/smiles features and term structure, and compare it relative to other equity markets and asset classes. For instance, a couple of measures we like to watch are the VIX/Skew (here) and the VIX/VXV (here) ratios.

Hence, in today’s article, we present the VIX/RVX, which measures the ratio between the implied volatility of the SP500 and the Russell 2000, a small-cap stock market index. As you may know, the ‘small cap premium’ has been a crowded study in the empirical academic research, which started from the early work of Rolf Banz (1981) who founded that ‘smaller firms have had higher risk-adjust returns, on average, than larger firms’. Then, in their paper The Cross-Section of Expected Stock Returns (1992), Fama and French found that value and small cap stocks, on average, outperform growth and large carp stocks. As you can see it on the chart, an interesting development has occurred over the past few days following the huge spike in volatility. The VIX/RVX, which has constantly been above parity since 2006, is now sitting at 0.83. In other words, according to the index, the Russell 2000 equity market carries less risk than the SP500. The question now is: what explains this sudden drop in the ratio?

If we look at the week-on-week change in both indexes, we can first notice that, at current levels, the WoW change of 11.6 in the VIX came in at 5th position in the index history, just a 0.3 ‘shy’ of the October 1997 move (here). However, if we now look at the change in the implied volatility of the small caps, the RVX index barely changed (+2.3) over the past week, meaning that the drop in the ratio was only coming from the VIX move (here).

Hence, this leads us to an interesting conclusion: it seems that there is much more financialization going on with the VIX than with the RVX, either through the creation of single and double-levered long and short VIX ETFs products, or from a volatility-targeting and risk-parity perspectives (are those strategies more oriented towards the SP500?).

Chart: Relative Implied Volatility – VIX / RBX ratio (Source: Eikon Reuters)

Great Chart: ECB Total Assets vs. EuroStoxx50 (18M Lag) – Yearly Change

One of the main topics of the year is the central banks’ balance sheet unwind, and the potential effect it can have on asset prices. As JP Morgan (and other sell-side institutions) pointed out, if we look at the annualized monthly net bond flows, the top 4 central banks (Fed, ECB, BoE and BoJ) will switch to net sellers in October 2018 (here). BNP Paribas published an interesting chart lately of the weighted average 10-year G4 bond yield overlaid with the G4 monthly bond purchases (here); we can clearly see that the increase in the total purchases has helped to push overall 10Y yields on the downside since 2010, hence eased financial conditions and stimulated the refinancing activity. However, what will happen to LT yields now that the purchases are expected to fall in 2018?

Many market participants have argued that the constant increase in central banks’ balance sheet has levitated all asset classes, and particularly the stock market; therefore, one economic area we are watching closely during the unwind is the Euro zone. If we look back three years ago, when Mario Draghi announced the launch of the 60-billion Euro bond-buying program on January 22nd, 2015, the ECB balance sheet was totaling 2.15tr Euros and the equity market EuroStoxx50 was trading at 3,400. As of today, the central bank’s assets are north 2.3tr Euro (the ECB balance sheet surpassed the Fed’s one last summer and is now worth 4.5tr Euros), while the EuroStoxx50 Index is up a mere 200pts, currently trading at 3,600 (here). We can clearly notice that the ECB effect on European equities was non-existent. It looks like the European equity market has been a dead market over the past couple of years; the Eurostoxx 50 has been trading sideways within an 800-point range between 2,900 and 3,700 and sits at its 50% Fibonacci retracement from its mid-June-2007 peak to Feb-2009 trough.

Hence, we chose this week to overlay the yearly change in the ECB balance sheet’s total assets with the yearly change in the equity market (18-month lag). As you can see, the two times series have shown some co-movements since the Great Financial Crisis; a decrease in the ECB assets is usually associated with a negative YoY performance in the EuroStoxx50 18 months later. For instance, the ECB balance sheet yearly change switched from +60% in June 2012 to -24% in January 2014 amid early LTROs reimbursement by European banks. If we look at the lagged performance of the equity market, the yearly change in the EuroStoxx50 index went from +20% in the summer of 2012 to -18% in June 2014.

In October 2017, the ECB cut its bond-buying program to 30bn Euros a month starting January 2018 for a period of 9 months, and the market expects that the central bank will taper QE to final three months of the year. With the yearly change on the ECB assets starting its downward trend, our question is the following: will the growth and investment story in the Euro area offset the expected downturn in equities?

Chart: ECB Total Assets vs. EuroStoxx50 (18M Lag) – Yearly Change (Source: Reuters Eikon)

ECB vs Asset.png

Great Chart: GBPUSD vs. FTSE 100

As we are closely approaching our 1.40 target for Cable (here), we chose an interesting chart this week that shows a scatter plot of the UK equity market (FTSE 100) with GBPUSD exchange rate, using a weekly frequency since January 2009.  Even though the relationship is not as clear as for Japan Equities and USDJPY (here), we can still observe a negative ‘Pavlovian‘ relationship where a cheaper currency usually implies higher equities. For instance, the British pound was massively sold post referendum (June 2016) on the back of an elevated political and economic uncertainty, high volatility and negative investors’ sentiment. Cable plunged from 1.44 a week before Brexit vote to reach a low of roughly 1.20 in October 2016 before starting its recovery in the first quarter of 2017.

One interesting observation is in the equity market; even though the FTSE 100 sold from 7,000 in April 2015 to 5,700 in February 2016 prior the event (as Cable), the post-Brexit rounds of Sterling depreciation played in favor of UK equities. However, over the past few months, the situation recovered in the UK, both the uncertainty and the volatility eased. If we look at the Economic and Political Uncertainty index, a monthly series based on newspaper coverage developed by Baker, Bloom and Davis, it is down from almost 1,200 (summer 2016) to 200, its prior Brexit average, bringing Cable’s 1M ATM implied volatility from 19 to 7.85 (here). At the same time, the 3-month 25 Delta Risk Reversal is back into the positive territory (from -6 in June / July 2016), meaning that the implied volatility on calls is more expensive than puts (here).

With an equity market closing at 7,730 on Friday and Cable at 1.3850 (flirting with the 1.39), we are curious to see if the relationship will continue this year. Hence the question is: will the Footsie break its 8,000 psychological resistance while Cable continues its momentum?

Our view is that the Bank of England may surprise the market in 2018 concerning its interest rate path. With the December 2019 short-sterling futures contract trading at 98.88 (i.e. implied rate of 1.12 by the end of 2019), market participants are currently pricing in two hikes for the next couple of years. We think that three to four hikes is more appropriate to the current economic climate, and policymakers may send a signal in the February update of its inflation forecasts, triggering some moves in the short-term interest rate market. We think that a potential move in the forward IR curves will benefit to the Sterling pound, however equities may struggle to reach new highs and break above the 8,000 level.

Chart: Cable vs. FTSE 100 (Source: Reuters Eikon)

Great Chart: Gold vs. US 5Y Real Yield

We showed in many of our charts that 2017 was the year where some of the strong correlations between assets classes broke down. We showed USDJPY vs. TOPIX (here, here), Cable (here) and EURUSD (here) vs. the 2Y and 10Y interest rate differentials, and this week we chose to overlay Gold prices with 5Y US real interest rates. As we explained it in our study on Gold (here), the relationship between Gold and US [real] rates is easy to understand. The precious metal is a non interest-bearing asset, meaning that a typical investor doesn’t get any cash-flow from owning it (unlike dividends for stocks and coupons for bonds), and has usually a storage cost associated with it. Therefore, the forward curve of the ‘currency of the last resort’ (Jeffrey Currie) is usually upward sloping, in other words Gold market is in contango, with the forward price equal to the following:

Reg.PNG

Hence, if real interest rates start to rise, a rational investor would prefer to reallocate his wealth to either US Treasuries or Treasury Inflation Protected Securities (TIPS) and receive coupons rather than keeping a long position in a commodity that has a ‘negative carry’.

As you can see it on the chart, Gold prices (in US Dollars) and the 5Y TIPS real yield have shown some strong co-movements over the past 5 years, until the summer of 2017 when the two times series diverged. If we would follow recent moves on the market, the late surge in Gold prices (currently trading at 1,340 $/ounce) would imply a 50 to 60 bps decrease in US real interest rates (note that if we regress the change in Gold prices on the change in the 5Y real yield using weekly data since 2013, we find that a 1% increase in real yields lead to an 8.7% depreciation in Gold prices). And lower real rates would either come from higher inflation expectations or lower nominal interest rates. With the 5Y5Y forward inflation swap currently trading at 2.11% and up 30bps over the past 6 months, core inflation and core PCE YoY rates at 1.8% and 1.5% slightly moving to the upside, and oil prices still trending higher with WTI front month contract trading at $64.5, there is room for higher inflation prints coming ahead. However, if the two curves were to converge in the short term, the [sharp] move would come from either [lower] Gold prices or [lower] Treasury rates.

Our view is that the divergence will persist in the beginning of 2018, with inflation remaining steady / slightly increasing and US interest rates failing to break new highs on the long end of the curve (5Y and 10Y). The main reason for that is that we think market’s confidence on the Fed’s 4 or plus hikes will slow down in the coming months on the back of lower-than expected fundamental, depriving the yield curve from steepening too much.

Chart: Gold prices vs. US 5Y TIPS (inv.) (Source: Reuters Eikon) 

WeeklyGold.PNG

 

Great Chart: TOPIX vs. USDJPY

As we always like to look at the Japanese Yen charts (USDJPY, AUDJPY, MXNJPY) as a sort of alternative barometer of investors sentiment and overall financial conditions, we chose an interesting chart this week that shows a scatter plot of the Japanese equity market (TOPIX) with USDJPY. The two assets have shown a significant relationship over the years, especially since Abe took office in Q4 2012 and the BoJ introduced QQME (i.e. extremely accommodative monetary policy) on April 3rd 2013. Investor Kyle Bass was one of the first to introduce the term Pavlovian response to this ‘weaker yen, higher equities’ relationship in Japan, which brought a lot of ‘macro tourists’ instead of long-term investors.

However, we noticed that the relationship between the Yen and the TOPIX broke down in Q2 2017. While the Japanese equity market has continued to soar over the past few months, currently flirting with the 1,900 psychological level (its highest level since 1991), USDJPY has been less trendy and has been ranging between 107 and 114 (see divergence here). Hence, we decided to plot a scatter chart between the two assets using a weekly frequency since 2001.

As you can see, a strong Japanese Yen (i.e. USDJPY below 100) usually goes in pair with a weak equity market. For instance, we barely see the TOPIX index above 1,000 when the USDJPY trades below the psychological 100 level. However, as the exchange rate increases, we see more dispersion around the upward sloping linear trend; for a spot rate of 120, we had times when the TOPIX was trading at 800 and other times when it was trading at 1,800. We did a simple exercise and regress the exchange rate returns on the equity returns (both log terms) to see if we get some significant results, using the following equation:

As you can see, the coefficient Beta is economically and statistically significant at a 1-percent level. Using 16 years of data, we find that a 1-percent increase in USDJPY spot rate is associated with a 0.76% increase in the stock market.

We highlighted the point where we currently are in the chart (Today), which is a TOPIX at 1,889, its highest level in the sample, for a USDJPY spot rate of 112.80. We can notice that the point is located at an extreme level of dispersion, and the question we raised a few weeks ago was ‘Can the divergence between the equity index and the exchange rate continue for a while?’

We think that the stock market in Japan will struggle to reach new highs and generate some potential interesting returns in the months to come due to the poor performance of the banking system (strong weigh in the index) and the constant decrease in the effectiveness of the BoJ policy measures. We mentioned a month ago that the Japanese Yen was 26% ‘undervalued’ relative to its 23Y average value of 99.3 according to the Real Effective Exchange Rate (REER valuation) (see here), hence we find it difficult to imagine a super bear JPY / Bull TOPIX scenario. In addition, we also raised the fact that the current level of oil prices were going to deteriorate Japan Trade Balance in the future (see here), pushing back the current account in the negative territory and potentially impacting the stock market.

Chart: Scatter plot of TOPIX vs. USDJPY – weekly frequency (Source: Reuters Eikon)